Alasdair Allan

@aallan
4.9K Followers
337 Following
1.5K Posts
Scientist, author, hacker, maker, and journalist. Writes, speaks, and builds. An accidental privacy advocate. You can reach me at đź“« [email protected]. #MachineLearning #TinyML #ArtificialIntelligence #AI #LLM #EdgeComputing #Embedded #IoT #Mobile #Apple #BigData #Space #SmallSat #CubeSat
WEBSITEhttps://alasdairallan.com
BLUESKYhttps://bsky.app/profile/alasdairallan.com
MEDIUMhttps://aallan.medium.com
"It Wasn't Built for This," by me for Negroni Venture Studios. There were 1 billion commits on GitHub in all of 2025. The platform is now processing 275 million per week. GitHub Actions went from 500 million minutes per week in 2023, to 2.1 billion in a single week in April. Same number of developers, running agents that each behave like a hundred of them. #GitHub #AI https://negroniventurestudios.com/2026/05/27/it-wasnt-built-for-this/
It wasn’t built for this

GitHub’s platform uptime over the last 90 days is 86.98%. That is not a typo. Marek Šuppa, who built an independent tracker because GitHub stopped publishing aggregate uptime numbers on its own status page, has reconstructed the figure from GitHub’s incident feeds. IncidentHub counted 257 separate incidents between May 2025 and April 2026, 48 of […]

Negroni Venture Studios
Announcing https://agentlanguages.dev, a community-edited catalogue of the 21 programming languages designed for AI agents to emerge since late 2025. The most interesting thing to emerge? They fall into three philosophical camps — Syntactic, Verification, Orchestration — that disagree on the underlying problem. Genesis essay on the taxonomy: https://negroniventurestudios.com/2026/05/20/three-camps-alike-in-dignity/ #AI #LLM #ProgrammingLanguages
agentlanguages.dev — Programming languages designed for AI agents to write

A community-edited catalogue of programming languages designed for AI agents to author code, organised around three philosophical camps: syntactic, verification, and orchestration.

"Three camps alike in dignity," by me for Negroni Venture Studios. In the last six months, seventeen programming languages designed for AI agents to write rather than humans have appeared independently. They divide into three camps: syntactic, verification, orchestration. I built one of them (Vera), went looking for the others, and found a field. #AI #ProgrammingLanguages #Vera #FormalVerification #LLM https://negroniventurestudios.com/2026/05/20/three-camps-alike-in-dignity/
Three camps alike in dignity

In the last six months, programming languages designed for models rather than humans have started to appear. Work on most began entirely independently, initiated by people who had never heard of each other, but all roughly during the same window of time. I know this, because I built one of them. Vera is a statically […]

Negroni Venture Studios
"Documentation-First Programming," by me for Negroni Venture Studios, If AI writes the code, what does the human write? The part of the job most developers hate has now become the whole job. #LLM #AI #Documentation https://negroniventurestudios.com/2026/05/19/documentation-first-programming/
Documentation-first programming

I find it somewhat ironic that the part of the job that most developers hate is about to become the whole job. Because, at least as far as I can figure out, we are now in the era of documentation-first programming. Act I: The specification is the program A few months ago I started building […]

Negroni Venture Studios

I expanded the #Vera benchmark to six models across three providers (Anthropic, OpenAI, Moonshot). Kimi K2.5 got every Vera problem right while only managing 86% in Python and 91% in TypeScript. Across the flagship models, Vera and Python are level at 93%.

Yet there is no Vera in any training data, no examples on GitHub or Stack Overflow. Every token generated from a single spec document in the prompt. Language design doing a lot of heavy lifting.

https://veralang.dev

#AI #LLM #Programming

That's been a productive morning! https://github.com/aallan/vera/pull/415
The #Vera v0.0.101 release makes LLM calls first-class algebraic effects. An agent calling another model declares <Inference> in the signature. Type-checked, contract-verified, impossible to call from pure functions. Postconditions constrain model output at runtime. Empty string? Contract violation, not silent garbage. Compose <Http, Inference> to fetch content, then reason. Every side effect declared. As far as I know, no other language can do this! https://veralang.dev #AI #LLM #Programming
I built a programming language you're not supposed to read, https://veralang.dev. Vera is designed for machines to write. No variable names — typed De Bruijn indices instead. No style choices. Every function contract verified by Z3 SMT solver. Compiler errors are fix instructions for the model, not diagnostics for you. Pure by default, all effects explicitly declared and tracked. Point a model at it. #AI #LLM #OpenSource #Agent
https://negroniventurestudios.com/2026/02/28/a-language-designed-for-machines-to-write/
Vera — A language designed for machines to write

A programming language where verification is a first-class citizen. Mandatory contracts, algebraic effects, typed slot references. Compiles to WebAssembly — runs at the CLI or in any browser.

First signal for the end of what I’ve always called “the platform problem” for #AI. It’s not about the platform anymore, it’s about actions. It’s about AI acting, not chatting. https://venturebeat.com/technology/openais-acquisition-of-openclaw-signals-the-beginning-of-the-end-of-the
"The web just got a second audience," by me for Negroni Venture Studios. Google shipped WebMCP in Chrome. Cloudflare shipped Markdown for Agents. Both quiet announcements. Neither made the front page. The web just forked its audience, and the infrastructure is already being rebuilt to serve both halves. #AI #MCP #WebMCP https://negroniventurestudios.com/2026/02/18/the-web-just-got-a-second-audience/